ram-lexsi/aligntune-testrun-RAFT

TEXT GENERATIONPricing:Input $0.04 / Cached $0.008 / Output $0.08Concurrent Unit Cost:1Model Size:0.5BQuant:BF16Context Size:32kTool Calling:SupportedPublished:Aug 26, 2026Architecture:Transformer Featherless Exclusive Cold

The ram-lexsi/aligntune-testrun-RAFT model is a 0.5 billion parameter causal language model developed by ram-lexsi, fine-tuned from Qwen/Qwen2.5-0.5B-Instruct. It utilizes the RAFT algorithm and TRL backend, built with the AlignTune framework. This model is designed for general text generation tasks, leveraging its compact size and 32768-token context length for efficient deployment.

Loading preview...

Overview

The ram-lexsi/aligntune-testrun-RAFT is a compact 0.5 billion parameter causal language model, fine-tuned from the Qwen/Qwen2.5-0.5B-Instruct base model. Developed by ram-lexsi, it is built using the AlignTune framework, which supports various open-source models, algorithms, and backends.

Key Characteristics

  • Base Model: Fine-tuned from Qwen/Qwen2.5-0.5B-Instruct.
  • Fine-tuning Algorithm: Employs the RAFT (Retrieval Augmented Fine-Tuning) algorithm.
  • Backend: Utilizes the TRL (Transformer Reinforcement Learning) backend for training.
  • Context Length: Supports a substantial context window of 32768 tokens, allowing for processing longer inputs.
  • Framework: Developed within the flexible AlignTune ecosystem, enabling broad compatibility and customization.

Usage

This model is suitable for integration into applications requiring a small yet capable language model, particularly for tasks where the RAFT algorithm's benefits in fine-tuning are advantageous. Its 0.5B parameters make it efficient for deployment in resource-constrained environments, while the large context window supports complex conversational or document-based applications.